The fields of pharmaceutical development and therapeutic application both face substantial challenges. The therapeutic domain calls for more treatment alternatives, while numerous promising pre-clinical drugs have failed in clinical trials. One of the reasons is the inadequacy of Cross-drug Response Evaluation (CRE) during the late stage of drug development. Although in-silico CRE models offer a solution to this problem, existing methodologies are either limited to early development stages or lack the capacity for a comprehensive CRE analysis. Herein, we introduce a novel computational model named DeepCRE and present the potential of DeepCRE in advancing therapeutic discovery and development. DeepCRE outperforms the existing best models by achieving an average performance improvement of 17.7\% in patient-level CRE and a 5-fold increase in indication-level CRE. Furthermore, DeepCRE has identified six drug candidates that show significantly greater effectiveness than a comparator set of two approved drugs in 5/8 colorectal cancer (CRC) organoids. This highlights DeepCRE's ability to identify a collection of drug candidates with superior therapeutic effects, underscoring its potential to revolutionize the field of therapeutic development.